What are the verified coding and reasoning benchmarks for Gemini 3.7 Flash?

Gemini 3.7 Flash achieves SWE-bench Verified: 67.1% (Thinking (Low)), LiveCodeBench: 51.9%, AIME: 81.5%. Verified sources: Gemini 3.7 Flash Technical Report (Google DeepMind).

How does Gemini 3.7 Flash compare to rival frontier coding models on SWE-bench and pricing?

Gemini 3.7 Flash achieves SWE-bench Verified: 67.1% at $0.15/$0.6 per 1M tokens. Direct peer comparison: North Mini Code (53% SWE-bench, $0/$0); DeepSeek V4.1 Flash (56.2% SWE-bench, $0.3/$0.9); Llama 4 Scout (69% SWE-bench, $0.6/$2.4); Gemini 3.8 Flash (68.4% SWE-bench, $0.15/$0.6).

Does Swarm isolate Gemini 3.7 Flash code edits from other agents?

Yes. Swarm provisions isolated Git worktrees for every session and subagent, preventing file race conditions and keeping your main working tree clean until verification passes.

Can Gemini 3.7 Flash be assigned to specific subagent roles?

Yes. In ~/.swarm/config.json, you can configure Gemini 3.7 Flash as Primary/Planner (DAG task decomposition), Coder (implementation in isolated worktrees), or Finder (read-only symbol search).

Gemini 3.7 Flash — Specs, Context & Capabilities

Q: What is the recommended Swarm multi-agent topology for Gemini 3.7 Flash?

Swarm recommends Gemini 3.7 Flash as Coder Subagent (Tier 3: High-Throughput Subagent Engine). High execution velocity and low pricing enable deploying parallel workers across separate worktrees. Best paired with: Claude 3.7 Sonnet as Lead Planner.

Q: How do I run Gemini 3.7 Flash with Swarm?

Configure Gemini 3.7 Flash via CLI using: swarm ctl model set --provider google --model google/gemini-3.7-flash, or switch models interactively inside the terminal TUI with /models.

Gemini 3.7 Flash Benchmarks & Agentic Accuracy

Verified primary research & developer preference metrics

Primary Sources Verified

SWE-bench Verified

67.1%

Thinking (Low)Google DeepMind Report ↗

LiveCodeBench

51.9%

ThinkingLiveCodeBench ↗

LMSYS Coding Elo

1370Elo

Coding ArenaLMSYS Arena ↗

AIME 2024 / 2025

81.5%

ThinkingGoogle DeepMind Report ↗

Pricing ($ / 1M tokens)

$0.15/ $0.6

$0.03 cached readStandard API Pricing

What This Means in Real Software Development

67.1% on SWE-bench Verified.
Large 1M token context window processes full build logs, unit test failures, and AST symbol tables simultaneously.
Scores 81.5% on AIME 2024, proving strong deductive logic for complex logic fixes.
Extremely economical $0.15/M input pricing.

Swarm Multi-Agent Topology

Tier 3: High-Throughput Subagent EngineRole: Coder Subagent

Best Paired With: Claude 3.7 Sonnet as Lead Planner

High execution velocity and low pricing enable deploying parallel workers across separate worktrees.

Git Worktree Isolation Advantage:

Runs build, lint, and test validation across isolated Git worktrees concurrently.

What Developers & Practitioners Are Saying

Real-world field observations, production strengths, and developer consensus

Field Consensus

"Predecessor to 3.8 Flash, still widely used for high-throughput API endpoints and batch summarization."

Verified Real-World Strengths

Fast inference speed
Very low operational costs
Solid general code comprehension

Production Caveats & Gaps

Slightly lower tool-calling precision than 3.8 Flash

Primary Sources & Methodology Citations

Gemini 3.7 Flash Technical Report — Google DeepMind (January 2026)

Head-to-Head Frontier Comparisons: Gemini 3.7 Flash vs Rivals

Standardized, side-by-side benchmark & economics comparison against peer models

Flagship Comparative Matrix

Model	Role / Tier	SWE-bench	LiveCodeBench	LMSYS Elo	AIME	Input / Output
Gemini 3.7 FlashGoogle GeminiCurrent	Tier 3: High-Throughput Subagent Engine	67.1%	51.9%	1370	81.5%	$0.15 / $0.6
North Mini CodeIndependent Compare Gemini 3.7 Flash vs North Mini Code →	Tier 3: Zero-Cost Reflex Specialist	53%	46%	—	—	$0 / $0
DeepSeek V4.1 FlashDeepSeek Compare Gemini 3.7 Flash vs DeepSeek V4.1 Flash →	Tier 3: Refined Open Subagent	56.2%	52%	1360	74.5%	$0.3 / $0.9
Llama 4 ScoutMeta Compare Gemini 3.7 Flash vs Llama 4 Scout →	Tier 2: Global Context Specialist	69%	61%	1380	—	$0.6 / $2.4
Gemini 3.8 FlashGoogle Compare Gemini 3.7 Flash vs Gemini 3.8 Flash →	Tier 3: High-Throughput Subagent Engine	68.4%	53.8%	1380	84%	$0.15 / $0.6

Explore All AI Model Comparisons Hub (Gemini 3.7 Flash & Flagship Peers) →

Gemini 3.7 Flash model overview

Gemini 3.7 Flash is listed in the public Swarm catalog through Google Gemini. This page summarizes the verified catalog fields used by Swarm without claiming real-time access for every provider account.

Provider: Google Gemini
Family: Gemini
Context window: 1.05M tokens
Thinking support: low · medium · high
Availability status: Deprecated
Catalog ID: google/gemini-3.7-flash

Audio input
File input
Image input
Text input
Video input
Text output
Reasoning
Tools
Structured output
JSON
Batch
Deprecated

Run Gemini 3.7 Flash in Swarm

Verified CLI workflow & multi-agent architecture

Quickstart Guide →

bash — swarm run

$swarm ctl model set --provider google --model google/gemini-3.7-flash

# Or switch interactively inside the terminal TUI:

>/models

Subagent Role Matrix

Assign Gemini 3.7 Flash to specialized system roles in ~/.swarm/config.json:

Primary / Planner: Orchestrates full task DAGs and multi-stage pipelines.
Coder Subagent: Generates code and unit tests inside private Git worktrees.
Finder Subagent: Performs read-only code search, symbol discovery, and mapping.

System Agent Roles →Why Swarm Architecture →